Pain is frequent in diabetic neuropathy and is very hard to manage. Antiepileptic drugs have been used in treating pain for several decades. Their effectiveness has been described in different types of neuropathic pain, but when used as analgesics in painful diabetic neuropathy it still remains controversial. To clarify this effectiveness, a meta-analysis was performed to determine which antiepileptic drug had the best analgesic potential for managing pain in patients suffering from painful diabetic neuropathy. The search covered the Cochrane, MEDLINE, EMBASE, and LILACS databases, between January 1966 and September 2005. The following information was obtained from each article: criteria for diagnosing diabetic neuropathy, patients' age average, antiepileptic drug received and dose, sample size, duration of the disease and treatment follow-up, outcome measurement, evaluation of pain, and rescue medication. A combined 2.33 relative risk (95% confidence interval [CI] 1.88-2.88) was obtained; this result indicated that the antiepileptic drugs studied were effective for controlling pain in diabetic neuropathy. The corresponding necessary number to treat (NNT) values were established for evaluating which antiepileptic drug was most effective as an analgesic, according to our interests; pregabalin was shown to be the antiepileptic drug having the lowest NNT (NNT = 3.24 and 95% CI 2.12-6.81) for achieving greater than 50% analgesia in patients suffering from painful diabetic neuropathy. Antiepileptic drugs are frequently used in the specific case of diabetic neuropathy; the combined result of this meta-analysis has demonstrated their analgesic benefit. © 2007 U.S. Cancer Pain Relief Committee.